296_B34_Petitniot

07.05.2002

7:19 Uhr

Seite 296

B 3.4

EXPERIMENTAL ASSESSMENT AND FURTHER

DEVELOPMENT OF AMPLIFIED PIEZO ACTUATORS
FOR ACTIVE FLAP DEVICES

(*)

J-L. Petitniot, H-M. des Rochettes, P. Leconte

(*)
ONERA, Chtillon and Lille, France

Abstract:
RPA (Rotor Pales Actives) Franco-German project was launched three years ago to study the possible benefits
of implementing active trailing edge flaps on a helicopter main rotor. The main expected effects concern the
decrease of BVI noise in descent flight and the improvement of the dynamic behaviour of the rotor throughout
the largest possible flight domain. The technological solution adopted to deflect the flap of an 1/3rd scale rotor
demonstrator, uses an off-the-shelf elliptic amplified actuator, from Cedrat Recherche Company, driving an
innovative patented mechanism. The device has been tested as well on the BRAVoS hover rig to assess the
influence of centrifugal loads as in the S3 MA wind-tunnel to evaluate the maximum flap deflections with
respect to actuation frequencies and Mach number. Extrapolation to scale 1, which must take into account
distinct typical constraints, is not purely homothetic and requires further developments for the actuator which
are under way, especially in order to respect the mass specifications.
Introduction
Electro active materials and hybrid components
integrating them have been initially developed for
the needs of the navy detection (sonars), the medical
imaging and low speed sub-micro positioning. The
idea to use them as actuators for the dynamic shape
control of structures, with a notion of high
mechanical power capability, compared to hydraulic
devices, raised less than about a ten years ago. This
activity, strongly encouraged from its start by the
helicopter community, also begins to trigger some
interest for the control of vortices moving on small
lift surfaces of fixed wings in the airplanes, missiles
and drones domains. The advanced concepts of selfadaptive control of helicopter rotor blades, taking
benefit of the so-called smart materials, have been in
fact strongly considered to solve the following
problems of increasing difficulty: improvement of
the aerodynamic efficiency by delaying the stall
limits , reduction of vibrations to improve cabin
comfort and gun shoot accuracy, and reduction of
noises caused by particularly intense vortex-blades
interactions during descent flight. These electrically
driven actuators, requiring little space with a quite
interesting volumic energy ratio and having a short
response time, are attractive for the manufacture of
reduced scale demonstrator. At full scale, these clean
and compact devices, which can be largely
distributed are considered to be also able to
advantageously replace classical hydraulic or
electromagnetic solutions thanks to clearly higher
volumic and mass powers, admissible force densities
and operating frequencies, but some technical
limitations and onboard flight constraints
nevertheless exist. In the two cases, but much more

296

at full scale where the problem of mass limitation is

severe, due to the actual principle and energy storage
capability of these materials, energy recovery from
working structural parts can and will have to be
addressed. The technical specifications for reduced
and full scale are not basically the same and the
transposition of the devices from one scale to the
other is not immediate as far as the corresponding
technology doesnt necessary exist yet on the market
and as the development costs of such systems may
drastically become prohibitive moreover without the
guaranty of a satisfactory reliability.
The first results obtained in the second phase of the
RPA project on a wind tunnel scale rotor blade
section with an active flap, in order to validate the
predictions of the numerical simulations made in the
first phase, are already quite satisfactory even if
some improvements still have to be done. An offthe-shelf elliptic amplified actuator, from Cedrat
Recherche Company, was chosen for the active
system which was separately tested under
centrifugal and aerodynamic loads. A good behavior
under centrifugal loads was noticed. Performance
under aerodynamic loads, satisfactory at low Mach
numbers, remains to be improved at higher speeds.
If a weight excess is not too critical on a reduced
scale model in order to demonstrate the merits of
some advanced concept, even if it may influence the
desired effect, it cant be tolerated on a manned full
scale prototype. These requirements explain the
reasons of the studies which have been undertaken
in order to decrease the weight and consequently to
increase the mass energy efficiency of the existing
elliptic actuator, in two ways: an optimization of the
elliptic geometry which has allowed to obtain a

ACTUATOR 2002, 8th International Conference on New Actuators, 10 12 June 2002, Bremen, Germany

296_B34_Petitniot

07.05.2002

7:19 Uhr

Seite 297

noticeable gain and the use of composite material

which undoubtedly allows to reach another gap
order.
Actuators and electro active stacks
Solid state actuators have been first developed to
produce precise and micro positioning without great
energy and power requirements. Several American
and European companies propose low (200V) or/and
even high (1000V) voltages Direct Piezo Actuators
(DPA). They consist in multi-layer piezo ceramics,
pre-stressed by serial Belleville springs inside a
stainless cylindrical housing. The deformation is
typically about 0.1% and maximal displacement is
limited to 120 m. In counterpart, force capabilities
are naturally high , generally over 1 kN. These direct
linear actuators are adequate for quasi-static
applications but not for dynamic ones. In fact the
proposed driving electronics are current and power
limited and not optimized to control high electrical
capacitances. To overcome this limit, Cedrat has
developed Parallel Pre-stressed Actuators (PPA)
where the pre-stress is applied by an external
parallel spring. These actuators, lighter and without
moving part, can be more efficiently operated in
dynamic conditions.
PROPERTIES

Piezoelectric
Soft ceramics

Electrostrictive

Hard ceramics

T Curie

140 200C

320 360C

T service

70 100C

160 180C

5 45 C

Polarisation voltage
Polarisation duration
Operating Voltage
- according to polarity
- opposite to polarity
Stroke (for ' U= 100V)
Stroke stability with
frequency ( ' l<10% )
Stroke stability with
temperature ( ' l<25% )
Hysteresis
Modulus
Blocked stress
Operating stress (*)

+ 2 kV/mm

< 20 s
200 V max

+3 +4kV/mm

300 400 V max

0 - 100 Hz
5 45 C
45 60 GPa
100 150 MPa
75 100 MPa

6 10 F

2 3.5 F

70 100

500 1000

1.5 3 %

0.2 0.5 %

+50 + 60C

< +5C

Response time

(+)

The RPA project aims to evaluate the benefits of

active trailing edge flaps mounted on the blades of
the main rotor of a helicopter. Two concepts referred
to as direct-lift flap and servo-flap were at first
considered. The second using a small chord flap and
relying of aerodynamic pitching moment to induce
blade torsion is finally preferred.
R 2100
R 1 8 9 0 ( 0 .9 R )
R 1 4 5 1 ( 0 .6 9 R )

R 420
R 275

Fig.1: RPA blade dimensions and flaps localisation

The overall diameter of the Mach-scaled rotor is 4.2
m. The maximum blade chord is 140 mm. The flap
dimensions are 210 mm in span and 21 mm in chord.
Significant vibration and noise reductions have been
calculated with ONERA numerical simulation codes
and expected for three spanwise locations of the flap
(figure 1) .

< 5 s

Flap actuation solutions

0.056%

0 100C

Merit Factor

Heating /20C

150 V

0 - 1000 Hz

Capacitance ()
Loss factor tg G

The RPA Project

5 10 mn

- 20 V (static)
- 150 V (static)
0.050 0.060%
0.033%

> 15%
30 GPa
50 75 MPa
25 50 MPa

their low capacitance favourable to low current

demands, lower loss factors limiting heat dissipation
and higher blocked stresses allowing to support
higher external efforts,[2]. Their lack of deformation
can be fulfilled by a higher voltage range supply and
a possible operation in opposite polarity (Table 1).

< 2%
110 GPa

25 30 F
8 8.5 %
< 100 s

(*) for the whole stroke - () stacks dimensions : 10x10 mm2 - h=25mm
(+) for a continuous running at 100 Hz and for 'U= 200V

Table 1: Comparative properties of stack materials

Amplified Piezo Actuators (APAs), are long-stroke
linear actuators. Stacks are pre-stressed inside a steel
elliptic frame which produces a natural
amplification ratio of the displacement (between 2
and 5 according to the two axes ratio). This principle
procures higher deformation level (between 0.3 and
3 %) compare to first types of actuators with only a
slight decrease of the force capability.
All these actuators, which have to be much more
considered as low speed micro positionners, make
use of soft ceramics stacks because of their higher
deformation,[1]. For dynamic purposes, hard or halfhard ceramics would have to be preferred because of

A large array of solutions was considered during the

design phase: hydraulics, electric motors and various
smart materials systems such those used or
developed by different Universities in U.S (L-L
amplification or double X frame),[3]. It appears that
APAs manufactured by Cedrat gives the best results
in terms of specific energy. Moreover, the long axis
of the elliptic frame can be easily centred along the
25 % pitching axis and the naturally amplified and
linear trajectory of the short axis extremity at 90 is
favourable to drive the flap.

Fig.2 : APA 230

Fig.3 : APA500L

In a first step, APA230 type was selected and was

used for centrifugal tests. With the APA500L,
recently proposed, the efficiency is nearly doubled.
This new actuator, smaller and lighter was used for
aerodynamic wind tunnel tests.

ACTUATOR 2002, 8th International Conference on New Actuators, 10 12 June 2002, Bremen, Germany

297

296_B34_Petitniot

07.05.2002

7:20 Uhr

Seite 298

Centrifugal tests
A specific model was designed and built to evaluate
the capability of the APA230 to sustain the
centrifugal loads which it would have to cope with
on the future rotor model, [4]. The actuator is
clamped on a surrounding rectangular frame on one
side and is supported by a centrifugal blade on the
other side to withstand the torque generated by the
overhang under centrifugal field (figure 4).

Fig.5: centrifugal model

The flap deflection was recorded versus the

amplifier voltages at different actuation frequencies,
respectively at 0 g and 2000g. The dynamic
component was limited to 55 V because of the
amplifier current limitation. The stroke loss between
the two acceleration levels was less than 10% and a
good repeatability of measurements without too
much noise was observed (figure 7).
Aerodynamic qualification

Fig.4: active flap system

A 2D-blade section aluminium model incorporating

the whole active system and equipped with the
previous similar instrumentation has been
manufactured (figure 8), [4]. The selected OA312
airfoil with a 140 mm chord exactly matches the
future rotor blade model.

Fig.6: BRAVoS test rig

Fig.8: Flap active system

The flap is deflected by means of a lever blade. In

fact the flap rotates around a pseudo axis given by
the flexion deformation of two composite hinge
blades. At rest, i.e without power supplied, the flap
lies in the maximum downward deflection. The
whole mechanism equipped with a dummy flap,
mocking realistic mass and inertia, was installed
inside an aluminium frame, closed by trap doors in
order to avoid any aerodynamic influence (figure 5).
Several sensors recorded various parameters such
the flap rotation by two small Hall effect sensors, the
effort on the lever blade by a strain gauges bridge
and the internal temperature of the closed blade box.
The whole model was mounted for the rotary tests
atop the BRAVoS rig (figure 6).

Fig. 9: Model in S3MA

The tests were carried out in the S3MA blow-down

ONERA wind-tunnel located at Modane. A rotating
hydraulic jack connected to the left-hand side of the
model allows to perform steady and unsteady
pitching movements (figure 9).
M ACH 0 . 6 5

Fl ap angl e ver sus st at i c wi ng i nci dence

Fl ap angl e ( )
10

ps eudos t at i c at 0. 5H z

8
6
Right f lap an gle

Cen t r al f lap an gle

Lef t f lap an gle

0
-2

-2

-4
-6
-8
W i ng i nci dence ( )

G f l a p ()
4

Fig.10: Flap deflections at Mach = 0.6

G f l a p = f (U a ) open loop

U a =90+/ -55 Volt s

3
2
1
0
-1

2000g
f requency 1Hz

-2

f requency 20 Hz
-3

f requency 50 Hz
U a (volt )

f requency 80 Hz

-4
30

40

50

60

70

80

90

100

110

120

130

140

150

Fig. 7: Flap deflection versus voltage under 2000g

298

Tests were made at several Mach numbers (0.3,

0.45, 0.65 and 0.80) and several angles of attack. At
low Mach numbers, a fairly good agreement can be
seen between the three locations of flap deflections
measurements. At 0 = 0.6, (figure 10) the
discrepancies between lever blade and flap ends
measurements begin to increase and become
significant at M = 0.8 .
A comparison of the capabilities of the devices
studied at the University of Maryland and at
ONERA is given table 2. The use of composite
material for the actuator frame is going to allow to
multiply by about 2.5 the energy to weight ratio.

ACTUATOR 2002, 8th International Conference on New Actuators, 10 12 June 2002, Bremen, Germany

296_B34_Petitniot

07.05.2002

Actuator
technology
Voltage (V)

7:20 Uhr

Seite 299

CEDRAT
CEDRAT
Maryland Maryland
th
APA500L
APA230
1/7 scale
full scale
th
mass
blade
For 1/ 2.62
blade
production
model
model
prototype
mass production
fabricated
in-house
Multi-layer Piezostacks Piezostacks on Piezostacks in
elliptic housing elliptic housing
With L-L
actuator
mass
configuration double Lever
production
(8 layers) amplification
134 o
0 -120
0 - 200
0 - 200
400

Mass of actuator
14
634
(grams)
8
53 (*20 )
Blocked force (N)
165
2540 (*889 )
Total stroke ()
Stiffness
0.048
0.021
-6
(10 N/m)
Resonance
beyond 150 beyond 150
frequency (Hz)
Energy to
53.1
weight ratio of
23.6
(*28 )
Actuator
3
(x10 mN/kg)
Ratio of Energy..
0.44
1
/L-L Energy..
Width (mm)
33
71
(chord axis)
Length (mm)
60
183
(span axis)
2 to 5
19
Thickness (mm)
Peak-to-peak flap
11.5
11.5
deflexion
* test with spring applied on actuator

CEDRAT
APA500L
x 1.5
Theorical
Full scale
sca
Piezostacks
in elliptic
housing

CEDRAT
APA750
Delivered
Full scale
Piezostacks
in elliptic
housing
new model

0 - 150

0 - 170

250

208

700

600

800
230

570
500

1250
750

920
1150

3.48

1.140

1.67

0.80

800

450

300

218

184.0

342.5

334.8

440.8

3.47

6.45

6.31

8.30

69

55

82.5

60.5

140

145

217.5

213.7

10
4.5

10
15
6.7
9.5
(measured) (extrapolated)

15
10.3
(hoped)

Tableau 2: Capabilities of smart actuators

Full scale prototype
The transposition of the technical solution to full
scale has to take into account various aspects such as
the technological and onboard constraints. For
weight, technical and cost reasons, five APA750
(APA500 x scale 1.5) and not two actuators will
have to be used and a composite frame cant be
avoided (table 3), [5].
Model (1 / 2.619)

Specifications

APA 230

APA750
600

Mass of one actuator (g)

250

208

Specific Energy (J/g)

- actuator
- whole system

0.184
0.070

0.342
0.116

Volumic Energy (J/dm3)

0.476

0.894

2 300 g

Centrifugal acceleration
Admissible mass increase
Actuator Mass (g)
Total Mass (g)
Torsion inertia axis
Maximal Voltage

570

Located at 25% of the pitch axis

unlimited

++

Power

unlimited

Cost of an actuator ( x N)

+ 20 kF (X 2)

Running clearance
Max Flap turning
( in unloading condition)
Hinge moment
( blocked moment)
Bandwith
Life Duration

+8/-5 +10/-5

Environment

APA500 x 2.619
3640

+ Direct
not viable
l i
0.15 0.30 J/kg
++
800 +/-30 g

10 15 %
(350 550 grams)
500
500
416
654

Prototype Scale 1

APA500

0.2 1.6 m.N

realized: 0.50 with 2 actuators
80 Hz (5:)
5
++ 5 h (3. 10 cycles) 115 rd/s
Wind tunnel Conditions

++

First tests of filamentary wounding have been done

to manufacture the composite frame. A carbon fiber
with an intermediate Young modulus has been
chosen (figure 12). A fiber volume of 55% has been
obtained. The high specific rigidity is favourable to
an increase of the resonance frequency. The higher
strength of the carbon epoxy allows to increase the
pre-stress of the stacks and to reduce the thickness
of the frame. The high conductibility of carbon and
the low dilatation in the fiber direction are
interesting to improve the positionning accuracy.
Efforts still have to be borne on the stacks stand
zones.
Conclusion
First results obtained in the RPA project with the
combination of the existing Cedrat actuators and the
specific transmission mechanism are rather
encouraging and are very competitive compared to
other solutions studied elsewhere. Even if minor
modifications are still required on the system, the
centrifugal aspects were rather easily mastered. The
behavior under aerodynamic loads has still to be
better assessed, especially at high speeds. The
geometry and the manufacture of the carbon epoxy
composite frame has to be industrially validated.
Short term modifications which can allow to hope
substantial energy savings concern the nature and
the disposal of piezoelectric ceramics of the stacks
and a deep modification of the actuator frame
profile. Other further improvements will concern the
driving electronics and energy recovery as well from
the active systems as from the surrounding working
elastic structures featuring active components.

5 10 %
(2.5 5 kg)
3000 (1960)

7280 (4880)

5765 (4725) 10045 (7645)

Located at 25% of the pitch axis

270 V DC
200 2000 W by blade
total limit : 10 kW max
65 kF (x 5)

Prohibitive (x 2)

++
5 15
15 80 m.N
30 Hz

++

Infinite (sfail safes)

-50 80C 85%RH
saline corrosion

References
[1] B. Camus, J-L Petitniot, H.M des Rochettes
Matriaux et Actionneurs pour le contrle de
forme des structures, 3 Journes dEtudes SEE
et INSA, Insa de Lyon, 16-17 septembre 1999
[2] J-L. Petitniot, H.M des Rochettes,Materials and
actuators for the shape control of structures, 8th
Symp. SPIE on Smart Structures and Materials
Newport Beach, Ca, USA, 4-8th March 2001

- unfavourable , + favourable , ++ very

Table 3: Reduced scale and full scale specifications

The effectively delivered APA750 is in fact not
homothetic with the APA500 one (figure 11). A
thinner frame section along with a smaller short axis
have allowed to increase the efficiency.

Fig. 11: APA750X

Fig.12: Composite frame

[3] J-L. Petitniot, Rotor gouverne active Etude

de faisabilit technologique, R.T n98/22
DMSE, ONERA, juin 1998
[4] P. Leconte, H-M. des Rochettes, Experimental
assessment of an active flap device, AHS 58th
Annual Forum, Montreal, Canada, June 2002
[5] J-L. Petitniot, Etude conceptuelle de pale active
dhlicoptre Echelle 1 , R.T n99/69, DMSE,
ONERA, mars 2000

ACTUATOR 2002, 8th International Conference on New Actuators, 10 12 June 2002, Bremen, Germany

299